[HIVE-17921] Aggregation with struct in LLAP produces wrong result

• Type: Sub-task
• Status: Closed
• Priority: Blocker
• Resolution: Fixed
• Affects Version/s: 3.0.0
• Fix Version/s: 4.0.0-alpha-1
• Component/s: llap, Transactions
• Target Version/s: 3.0.0
• Assignee: Saurabh Seth
• Reporter: Eugene Koifman
• Created: 27/Oct/17 15:51
• Updated: 17/Nov/22 08:54
• Resolved: 10/Sep/18 20:40

Description

Consider

select ROW__ID, count(*) from over10k_orc_bucketed group by ROW__ID having count(*) > 1;

in acid_vectorization_original.q (available since HIVE-17458)

when run using TestMiniLlapCliDriver produces "NULL, N" where N varies from run to run.
The right answer is empty results set as can be seen by running

select ROW__ID, * from over10k_orc_bucketed where ROW__ID is null

in the same test.

This is with

set hive.vectorized.execution.enabled=true;
set hive.vectorized.row.identifier.enabled=true;

It fails with TestMiniLlapCliDriver but not TestMiniTezCliDriver. See acid_vectorization_original_tez.q which has identical query.